Cloud computing literacy is understanding that an online service runs on real computers, stored information, power supplies, cooling systems, trained workers, software and networks. A child opens a digital worksheet and sees an answer appear instantly. That simple interaction may involve a local device, a broadband connection, a remote application, one or more data centres, identity systems and records that need to remain accurate. Calling the service ‘the cloud’ does not make its infrastructure or responsibilities disappear.
Ministry of Education V3.0 needs a practical guide to the physical and institutional foundations of digital learning. This reference explains what data centres do, how storage differs from computation, why cooling and reliable power matter, how cloud services divide responsibility, why digital resilience requires backup and exit planning, and how schools can assess the resource costs and educational value of online tools. It includes a data-centre dependency map, a safe ten-step audit, a Power Usage Effectiveness calculation, Singapore Green Data Centre Roadmap context, fictional case studies and an assessment rubric.
This is an independent eduKateSG educational framework, not an official Singapore Ministry of Education data-centre curriculum or technical operations manual. Data centres are controlled professional facilities. Students must not enter service areas, investigate sensitive network locations, collect credentials, attempt cyberattacks or operate electrical or cooling equipment. All classroom systems are conceptual or fictional.
1. What Is a Data Centre?
A data centre is a facility designed to host computing, networking and storage equipment under suitable power, environmental, physical-security and operational conditions. Some organisations operate their own facilities; others obtain computing and related services from specialist providers. A cloud service may allocate computing resources dynamically, but the work still occurs on physical equipment.
The student should distinguish three different activities: storing information so it can be retrieved, computing to process instructions and data, and communicating to send information between locations and users. They depend on one another, but a faster internet connection cannot automatically cure an overloaded application server, and more storage does not guarantee faster calculations.
Singapore’s Infocomm Media Development Authority Green Data Centre Roadmap identifies data centres as foundational digital infrastructure. IMDA states that facilities are energy- and resource-intensive and seeks sustainable growth through efficiency, appropriate equipment and greener energy. This public policy is a useful contemporary case of how apparently weightless digital activity depends on scarce physical resources.
2. The Eight Rooms Behind an Online Learning App
| Room | Useful function | Educational question |
|---|---|---|
| Learner and device | Presents the interface and accepts input | Can this user access and use the material? |
| Connectivity | Moves requests and responses among systems | Is the network route available? |
| Application | Applies teaching logic, records or other rules | What service is actually being provided? |
| Computing | Processes instructions or AI requests | Which type of work needs processing capacity? |
| Storage | Retains permitted content, records and related data | Who owns, secures and can recover the information? |
| Facilities | Provide power, cooling, space and physical protection | What keeps the equipment running safely? |
| People and governance | Manage software, maintenance, access, service and compliance | Who is accountable when something goes wrong? |
| Return and exit | Preserve learning through failure, change and correction | Can users continue or move to a new service? |
These rooms are a teaching framework, not a claim that every provider has identical architecture. A service might use multiple facilities, content-delivery systems or hybrid computing arrangements. Without appropriate evidence, the class should not claim to know exactly where personal data or computation resides.
3. Ten Steps for a Data Centre and Cloud Dependency Audit
Step 1 — State the educational purpose
The question is not ‘Does the school have enough cloud?’ It is whether the student can reach a relevant lesson, receive feedback, preserve work and demonstrate new understanding. Define the capability before evaluating the technology. A more expensive online product may not repair the actual learning gap.
Step 2 — Trace the observed service
Describe the authorised student action: open a lesson, submit a response, receive feedback. Distinguish what the interface visibly does from what the underlying software or infrastructure may do. Do not invent technical details based on an app’s promotional description.
Step 3 — Separate the device, network, cloud and application
Use a fictional diagram to show that a local device can work while an app fails remotely, or a remote app can remain healthy while the local connection is unavailable. This helps learners report the observed issue more clearly and helps teachers choose an approved fallback.
Step 4 — Identify data and processing requirements
Does the task need simple file retrieval, live collaboration, video processing or complex model inference? These have different computational and network demands. A learner should not assume that all digital tasks impose equivalent resource costs.
Step 5 — Recognise physical dependencies
Facility equipment needs appropriate electrical supply, environmental control, network connectivity and professional maintenance. Heat generated by computing must be managed. These facts can be understood through public diagrams without requiring access to facility schematics or security details.
Step 6 — Clarify the shared responsibilities
A service provider may maintain parts of the infrastructure, while the school controls its users, content, permissions and lawful educational purpose. Contractual arrangements vary. The institution cannot assume the provider automatically takes responsibility for student consent, quality of lessons, assessment validity or every form of data recovery.
Step 7 — Check privacy, rights and security
Map what student information leaves the school, for what purpose, for how long and under which approved rules. Protect copyrights and learning-resource permissions. A school should not upload identifiable assignments to an unapproved AI provider merely to obtain a convenient summary.
Step 8 — Inspect reliability and service recovery
Ask how the provider reports interruptions, restores service and enables authorised access to necessary records. Do not confuse advertised availability with a guarantee that every learning session will succeed. A fallback learning plan should be independent enough to function during an outage.
Step 9 — Count total educational and resource costs
Subscription fees, devices, connectivity, teacher training, energy, support, replacement and switching costs all matter. A provider may improve energy efficiency while total digital demand rises. The educational benefit and environmental cost must be interpreted together without claiming one metric tells the entire story.
Step 10 — Test the learner’s return
After an approved intervention or platform change, check whether a student can perform the relevant skill independently. Logins, page views and model response speeds are supporting technical information; they are not proof of learning. The strongest system remains useful and recoverable when its technology changes.
4. Power Usage Effectiveness: A Useful But Narrow Ratio
Power Usage Effectiveness, commonly abbreviated PUE, is a data-centre energy-efficiency metric comparing total facility energy with the energy used by IT equipment over a consistent measurement period. A ratio closer to 1 generally indicates lower non-IT overhead under comparable measurement conditions. It does not measure whether the IT work is educationally useful or whether the electricity is low-carbon.
| Quantity | Meaning | Example |
|---|---|---|
| IT equipment energy | Energy used by relevant computing, storage and network equipment | 1,000 kWh in a fictional measurement period |
| Facility energy | Total energy used by the data-centre facility over the same boundary and period | 1,500 kWh |
| PUE | Facility energy divided by IT equipment energy | 1,500 ÷ 1,000 = 1.5 |
| Non-IT overhead | Energy within the facility boundary beyond the measured IT total | 500 kWh in the simplified example |
The classroom calculation is exact under its invented numbers and boundaries. Professional PUE reporting follows specific measurement practices and conditions. Two facilities should not be ranked purely by a single PUE number when measurement boundary, climate, IT utilisation and operating context differ.
5. Singapore’s Green Data Centre Roadmap
IMDA launched the Green Data Centre Roadmap in May 2024. It aims to support at least 300 megawatts of additional data-centre capacity in the near term and further growth associated with greener energy deployments. This is a policy ambition, not a statement that every megawatt has already been delivered or used.
The IMDA’s November 2025 explanation highlights cooling improvements, energy-efficient IT equipment, workload optimisation and sustainability standards. The educational point is that data-centre sustainability involves both facilities and the computing work placed within them.
IMDA also maintains a Green Data Centre Standard page describing a management framework for improving efficiency. Its Energy Efficiency Grant information, updated February 2026 gives an example of policy support for adopting efficient equipment. Current eligibility rules are for the relevant operators and businesses, not individual student grants.
6. Worked Example: Better PUE, Higher Total Energy
Fictional exercise. Facility A uses 1,000 kWh for IT equipment with PUE 1.6, so total facility energy is 1,600 kWh. Facility B uses 1,400 kWh for IT equipment with PUE 1.3, so its total is 1,820 kWh. B has the lower overhead ratio but consumes more total energy during the stated period because its IT workload is larger.
The comparison illustrates why a strong relative-efficiency figure is not automatically an absolute energy reduction. Before judging which service is preferable, students should ask whether facilities deliver comparable useful computation, whether the data refer to comparable periods and what energy sources are involved.
7. Worked Example: A Cloud Lesson Stops Responding
A fictional class can open an official public information website but cannot retrieve submissions from its approved homework platform. Students know the device and some network services are working, but cannot infer whether the issue is a provider outage, a school account problem, application error or another cause.
The correct school response is to use a verified reporting channel, preserve available work in approved systems and follow the authorised alternative learning route. Learners should not disable protection or share login credentials. Teachers later reconcile missed content and check actual understanding.
This connects to the Internet Infrastructure reference and Education Emergency Continuity reference. The distinct contribution of this article is the physical and operational side of remote computing.
8. Worked Example: School Data Migration
A fictional school wishes to replace its learning platform. Existing work includes files, comments, progress records and permissions. The provider offers an export, but the new system requires a different data format. An institution should test whether the necessary records can be interpreted correctly, which materials it may lawfully reuse and how access and privacy remain protected.
The transfer should be reviewed by authorised technical, educational and data-governance staff. A file labelled ‘export complete’ is not proof that each learner can recover their work or that content rights permit continued use. The Vendor Dependency and Contract Exit manual details the relevant institutional procurement and exit controls.
9. A Cloud Service Accountability Card
| Field | Question | Evidence |
|---|---|---|
| Learning purpose | What actual capability should improve? | Target skill and independent check |
| Workload | What processing, storage and network functions are needed? | High-level functional requirement |
| Responsible parties | What does the provider handle and what does the school handle? | Current contract and approved operating policy |
| Data and rights | What information and content can be processed? | Lawful purpose, access, licence and retention |
| Facility dependence | What physical resources and service conditions matter? | Public, non-sensitive specifications |
| Availability | What interruption and restoration conditions apply? | Approved support and continuity terms |
| Exit | Can work, records and teaching continue elsewhere? | Tested export and permitted alternative |
| Return | Did the technology improve learning and accessibility? | Independent task and user experience evidence |
10. Age-Appropriate Learning Progression
| Stage | Classroom activity | Safeguard |
|---|---|---|
| Primary 1–3 | Distinguish device, internet and remote service | No personal-account analysis |
| Primary 4–6 | Trace a fictional homework request to a remote computer | Teacher-selected diagram |
| Lower secondary | Compare computation, storage, networking and availability | No live infrastructure probing |
| Upper secondary | Calculate facility energy and PUE | Explicit measurement boundaries |
| Pre-university | Evaluate cloud governance, sustainability and continuity | Respect data and institutional authority |
| Adult learning | Use approved cloud services and understand shared responsibilities | Professional processes for real operations |
11. Forty-Minute Cloud Infrastructure Lesson
| Minutes | Activity | Student output |
|---|---|---|
| 0–7 | Define a digital learning outcome and trace its components | Eight-room service map |
| 7–15 | Separate data, computation, connectivity and storage | Correct technical functions |
| 15–23 | Calculate fictional PUE and total facility energy | Numbers with units and interpretation |
| 23–30 | Introduce service failure and data portability constraints | Fallback and institution role map |
| 30–36 | Compare learning benefit and resource cost | Balanced provisional decision |
| 36–40 | Apply framework to a different online service | Independent transfer response |
12. Performance Rubric
| Dimension | Developing | Proficient | Advanced |
|---|---|---|---|
| System model | Treats the cloud as physically intangible | Traces computing, storage, network and facility functions | Explains dependencies and service boundaries |
| Energy | Mistakes lower PUE for lower total use | Calculates PUE and facility energy correctly | Distinguishes efficiency, useful work and environmental footprint |
| Responsibility | Assumes provider handles everything | Separates supplier and school duties | Recognises governance, licences and access limits |
| Resilience | Assumes an outage ends learning | Identifies authorised fallback and recovery | Evaluates portability and verified restoration |
| Education value | Equates usage with progress | Uses an independent learning check | Revises service choice after measured results |
13. The Gold Standard
A data-centre-literate learner understands where the ‘cloud’ physically runs, why electricity and cooling are part of the internet economy, and how useful computing depends on accountable service design. They can interpret a basic energy-efficiency ratio without claiming that it alone measures environmental or educational quality.
The best education technology is not the most elaborate infrastructure a school can purchase. It is infrastructure that serves a clear learning purpose, protects people, permits verified independent progress and remains intelligible and recoverable when systems change.
Related References and Public Sources
- MOE V3.0 — Reference Library
- Reference 52 — Internet Infrastructure and Cyber Resilience
- Reference 42 — Electricity Systems Literacy
- Reference 12 — AI Approval Gate
- Reference 28 — Digital Privacy Literacy
- Reference 33 — Vendor Dependency and Exit
- IMDA — Green Data Centre Roadmap
- IMDA — Launch of Green Data Centre Roadmap (May 2024)
- IMDA — Advancing Green Technology for Data Centres (2025)
- IMDA — Green Data Centre Standard
- IMDA — Data Centre Energy Efficiency Grant (2026)